Acta Biomaterialia 6 (2010) 1625–1629

Contents lists available at ScienceDirect

Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat

High resolution transmission electron microscopy study of the hardening

mechanism through phase separation in a b-Ti–35Nb–7Zr–5Ta alloy
for implant applications
Conrado R.M. Afonso a,*, Peterson L. Ferrandini b, Antonio J. Ramirez a, Rubens Caram b
a
Brazilian Synchrotron Light Laboratory (LNLS), Campinas, SP 13083-970, Brazil
b
Department of Materials Engineering (Dema), State University of Campinas (Unicamp), CP 6122, Campinas, SP 13083-970, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: b-Ti alloys are highly attractive metallic materials for biomedical applications due to their high specific
Received 29 July 2009 strength, high corrosion resistance and excellent biocompatibility, including low elastic modulus. This
Received in revised form 23 October 2009 work aims to clarify the hardening mechanism of a b-Ti–Nb–Zr–Ta alloy using different characterization
Accepted 9 November 2009
techniques. Ingots (50 g) of Ti–35Nb–7Zr–5Ta (wt.%) alloy were arc furnace melted in an Ar(g) atmo-
Available online 12 November 2009
sphere, homogenized, hot rolled, solubilized and finally aged at several temperatures from 200 to
700 °C for 4 h. Microstructure characterization was performed using X-ray diffraction, optical micros-
Keywords:
copy, scanning and high resolution transmission electron microscopy (HR-TEM). The 4 h aging showed
b-Ti alloy
Hardening mechanism
that the highest hardness values were found when aged at 400 °C and the HR-TEM images confirmed
Phase separation splitting of spots on the Fourier space map, which indicated the presence of a coherent interface
Aging between separated phases (b and b0 ) and explains the hardening mechanism of the alloy. Through geo-
metric phase analysis analysis, using the HR-TEM image, the localized strain map showed 5–10 nm
domains of the b and b0 phases. The combination of suitable values of yield strength, hardness and
low Young’s modulus makes Ti–35Nb–7Zr–5Ta alloy suitable for medical applications as a metallic
orthopedic implant.
Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1. Introduction
b-Titanium alloys are highly attractive metallic materials for
biomedical applications due to their high specific strength, high
corrosion resistance and excellent biocompatibility, including
low elastic modulus. The mechanical properties of Ti alloys
can be tailor adjusted through compositional variation, thermo-
mechanical processing and microstructural control [1–3]. It is
well documented that precipitation of the isothermal x phase
can impair the mechanical properties of b-Ti alloys used in
orthopedic implant applications. A previous study with Ti–
35Nb–7Ta and Ti–35Nb–7Zr–5Ta alloys revealed a hardness in-
crease during short-term aging, with maximum hardness ob-
served for aging at around 400 °C for 4 h, after which X-ray
diffraction (XRD) analyses revealed only a bcc b phase [2]. It
has been reported that such a high content of b-stabilizer ele-
ments added to Ti can lead to phase separation [4,5], which is
not fully understood, nor is its effect on the mechanical proper-
ties of the alloy. In the case of a 50Ti–30Zr–10Ta–10Nb (at.%)
alloy studied by Yang and Zhang [4], a peak separation between
* Corresponding author. Tel.: +55 19 35183132; fax: +55 19 35121004.
E-mail address: cmoreira@lnls.br (C.R.M. Afonso).
what they called b-Ti (a = 3.31 Å) and b-Zr (a = 3.57 Å) bcc
phases was found by conventional XRD analysis. A hardness
peak was observed for this alloy when aged at 400 °C (heat
treated from 200 to 900 °C), as for the Ti–35Nb–7Zr–5Ta alloy,
and was attributed to an undetected a-Ti phase precipitation.
In the case of the Ti–35Nb–7Zr–5Ta alloy, which has been
extensively studied in the literature due to its low Young’s mod-
ulus and good biocompatibility [1,2,6], the b-Ti and Nb phases
both have bcc structure with very similar lattice parameters
(3.31 and 3.33 Å, respectively), leading to very small differences
in XRD spectra and transmission electron microscopy (TEM) se-
lected area electron diffraction (SAD) patterns, avoiding the dif-
ferentiation of both phases through conventional analysis
techniques [2]. Improvements in mechanical properties such as
hardness, tensile and proof strength and a low elastic modulus
are very important and desirable for orthopedic implant applica-
tions. Therefore, a better understanding of such a phase separa-
tion requires the use of advanced characterization techniques,
such as high resolution transmission electron microscopy (HR-
TEM) and high resolution X-ray diffraction (XRD). The aim of
this work was to clarify the phase separation phenomenon in
Ti–35Nb–7Zr–Ta alloy through HR-TEM and XRD using synchro-
tron radiation.

1742-7061/$ - see front matter Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actbio.2009.11.010
1626 C.R.M. Afonso et al. / Acta Biomaterialia 6 (2010) 1625–1629
2. Experimental
Ingots (50 g) of Ti–35Nb–7Zr–5Ta (wt.%) alloy were arc furnace
melted in an Ar(g) atmosphere as presented elsewhere [2]. The in-
gots were then hot rolled at 1000 °C several times with intermedi-
ate reheating. After rolling the solution treatment comprised
heating at 1000 °C for 35 min, followed by water cooling. The final
1.8 mm thick plates were cut into 15  20 mm samples and then
encapsulated in quartz, under an argon atmosphere and aged for
4 h at 200, 300, 400, 500, 600 and 700 °C. X-ray diffraction (XRD)
patterns obtained using monochromatic synchrotron radiation
(k = 1.54 Å, as for Cu Ka) were collected for solubilized and air
quenched samples at room temperature (Troom = 25 °C) and sam-
ples aged at 200, 400 and 600 °C for 4 h under vacuum. High reso-
lution XRD experiments were performed in the D10B-XPD
(polycrystalline X-ray diffraction) beamline at the Brazilian Syn-
chrotron Light Laboratory, using a double crystal Si(1 1 1) mono-
chromator. TEM samples were prepared using Ar+ ion milling and
HR-TEM and analysis was performed at 300 keV using a JEM
3010-URP coupled to energy-dispersive X-ray spectroscopy
(EDS). Geometric phase analysis (GPA) [6] was applied to HR-
TEM images in order to evaluate displacement fields and image
variation in the positions of lattice fringes by filtering the Fourier
space, and JEMS software [7] was used to simulate b-Ti phase crys-
talline structures.
3. Results and discussion
Fig. 1 shows the XRD pattern obtained using synchrotron radi-
ation from a Ti–35Nb–Ta–Zr alloy sample aged at 400 °C for 4 h.
At first glance it seems that only one b phase peak was identified
for each interplanar distance (or 2h). However, when the peak
was magnified, as shown in detail for the (1 0 1) peak in the right
upper corner, a peak split was observed, revealing b and b0 phase
separation. The nearly parallel beam optics and the monochro-
matic radiation of the synchrotron beamline utilized provided bet-
ter resolution than a conventional bench Ni-filtered Cu Ka XRD set-
up. In addition, the larger photon flux of the synchrotron beamline
provides a larger signal to noise ratio (SNR), which facilitates the
identification of small volume fractions of crystalline phases.
Fig. 2a and b shows TEM bright field (BF) images with heteroge-
neous diffraction contrast microstructure, alternating dark and
bright regions of 10 nm. Fig. 2a shows a micrograph of a grain
boundary, revealing the absence of a phase precipitation. Fig. 2b
presents the Ti–35Nb–7Zr–5Ta alloy along the [1 1 1]b zone axis
after solution treatment followed by water quenching and aging
β β'
β highlighted region), which present a ð1 0 1Þb
Intensity (a.u.)
400ºC / 4 h
37,5 38,0 38,5 39,0
30 35 40 45 50 55 60 65 70 75 80 85 90
2 θ (degrees)
Fig. 1. XRD obtained using synchrotron radiation of a sample aged at 400 °C for 4 h.
at 400 °C for 4 h, revealing contrast between the b and b0 phases
[3]. However, neither x nor a phases were identified after 4 h
aging at the same temperature, even by TEM. The selected area
electron diffraction (SAD) pattern showed spots along the [1 1 1]b
zone axis, indicating the existence of only a b phase. Fig. 3a shows
a HR-TEM image of an intragranular b grain oriented along the
[1 1 1]b zone axis from the sample Ti–35Nb–7Zr–5Ta alloy aged
at 400 °C for 4 h.
The fast Fourier transform (FFT) of the HR-TEM image in Fig. 3b
 reflection, supporting the coexistence
reveals splitting of the ð1 0 1Þ
of two phases (b and b0 with 10% shortening in the d spacings of
2.3 and 2.1 Å, respectively) with similar body-centered cubic (bcc)
and body-centered tetragonal (bct) crystalline structures for b and

b0 , respectively, presenting the [1 1 1]b // [1 1 1]b0 ; ð1 0 1Þb //
 0
ð1 0 1Þb orientation relationship, and slightly different chemical
composition. Energy-dispersive X-ray analysis (EDX) nano-analysis
with 5 nm electron probe revealed Nb enrichment in the dark re-
gions (b phase) and Ta–Zr enrichment in the bright regions (b0
phase) of Fig. 2b, indicating the occurrence of a phase separation
phenomenon.
Composition fluctuations and segregation of heavy elements
such as Ta and Zr induced lattice parameter distortions and, conse-
quently, splitting of the diffraction peaks, revealing the coexistence
of b and b0 phases, which has been corroborated through XRD and
HR-TEM analyses. The small ð1 0 1Þ  d spacing difference between
the b and b0 phases, evident by XRD and HR-TEM, supports the
coherent interface observed during the HR-TEM analyses and ex-
plains the dramatic hardness increase of the alloy after the studied
heat treatment [4,5]. Analysis of the HR-TEM image FFT along the
[1 1 1] zone axis revealed that the {1 1 0}b plane spots are not sim-
 spots, as indi-
ilar, a clear splitting appearing just for the ð1 0 1Þb
cated by arrows in Fig. 3b. It should be noted that the other
{1 1 0}b-family plane spots do not show the above mentioned
splitting, indicating that the appearing b0 phase has a crystal struc-
ture different from that of b (bcc-Ti). The XRD results were used to
determine the b and b0 phase lattice parameters. The b phase ma-
trix presents a bcc (b-Ti) structure with lattice parameter
ab = 3.3 Å. The appearing b0 phase presents a body-centered tetrag-
onal (bct) structure with lattice parameters ab0 = 3.3 Å and
cb0 = 3.0 Å (c/a = 1.1), which reveals a slight shortening of the c
parameter with respect to the cubic b phase. The crystallographic
data for both phases were used to simulate the SAD patterns using
JEMS software [7], as shown schematically in Fig. 3c.
 interplanar spacing strain mapping,
Fig. 4 presents the ð1 0 1Þb
using GPA [8], which reveals the predominantly undistorted b ma-
trix (strain around 0) and the shortened b0 phase domains (strain
around 10%). Fig. 4a presents the GPA ð1 0 1Þb  interplanar spacing
strain map combined with the original HR-TEM image shown in
Fig. 3a. Fig. 4b shows the GPA ð1 0 1Þb interplanar spacing strain
map, which reveals a large number of dislocations, the predomi-
nantly undistorted b matrix and the b0 phase domains (dotted line
 interplanar spacing
strain of about 10%. This reduction of approximately 10% in the
 plane spacing, which in fact corresponds to ð1 0 1Þb
ð1 0 1Þb  0 , shows
very good agreement with the XRD results. For the above pre-
sented GAP analysis the b phase region was used as the reference
lattice for the geometrical phase calculation. For the GPA process-
ing a Lorentzian mask was used with a half-width of 1/4  d(1 0 1)
in reciprocal space, resulting in two-dimensional chemical and
strain maps with 1.2 nm spatial resolution. The GPA processing
was performed comparatively using different reflections in the
Fourier space, such as ð1  1 0Þb and ð0 1 1Þb
 spots for a [1 1 1]b-ori-
ented specimen, revealing the more severe ð1 0 1Þb  interplanar
space strain, which is supported by the FFT presented in Fig. 3b.
Under specific aging conditions the Ti–35Nb–7Zr–5Ta alloy
underwent a coherent spinodal decomposition of the b phase into
 C.R.M. Afonso et al. / Acta Biomaterialia 6 (2010) 1625–1629 1627

Fig. 2. Ti–35Nb–7Zr–5Ta alloy aged at 400 °C for 4 h. (a) TEM bright field (BF) images showing the absence of a phase precipitation along the grain boundary. (b) Detail of the
grain interior at the [1 1 1]b zone axis revealing diffraction contrast between the b (dark) and b0 (bright) phases.

Fig. 3. Ti–35Nb–7Zr–5Ta alloy aged at 400 °C for 4 h. (a) HR-TEM image of the region shown in Fig. 2b along the [1 1 1]b zone axis. (b) The Fourier transform (FT) reveals
splitting of the (1 0 1) reflections (white arrows). (c) Schematic representation of the Fourier space map of the [1 1 1]b zone axis, considering separated phases b with cubic
(bcc) and b0 with tetragonal (bct) structure obtained through simulation using JEMSÒ software.

two solid solution phases with different compositions and lattice
parameters, resulting in a high density of dislocations and elastic
strain associated with the domains of Nb-rich b and Ta–Zr-rich b0
phases. The strain fluctuations associated with this phase separa-
tion have been mapped through GPA analysis with a space resolu-
tion of 1.2 nm, as presented in Fig. 4b. The Nb-rich b and Ta–Zr-rich
b0 phases presented 5–10 nm phase domains, as revealed by the
low resolution TEM images and the GPA analysis. Table 1 shows
the EDX chemical compositions obtained with a 5 nm probe of dif-
ferent phase separation domains and at a grain boundary, confirm-
ing the coexistence of Nb-rich b and Ta–Zr-rich b0 phases in the
microstructure. It was reported by Ferrandini et al. [2] that after
heat treatment at 600 °C for 4 h a precipitation was detected along
the b grain boundaries. EDX nano-sized probe analysis confirmed
that even after aging at 400 °C for 4 h, segregation occurred, lead-
ing to a greater Ti concentration at the grain boundaries, following
a precipitation along the boundaries, which was observed for aging
heat treatment at higher temperatures (>600 °C) for 4 h.
Adopting suitable combinations of time and temperature for
aging of the Ti–35Nb–7Zr–5Ta alloy the single solid solution phase
(b phase) suffered phase separation, decomposing into coherent
phases (b and b0 ) with slightly different compositions and crystal-
line structures. Distortions along the strained nanometric domains
of the b and b0 phases induced hardening of the crystalline
structure.
The occurrence of coherent interfaces between the separated
phases could lead to an increase in the strength of a material. This
phenomenon is well known in Al–Cu alloys, in which Guinier–
Preston (GP) zones were identified by means of X-ray scattering
as being responsible for hardening, with the GP1 zones formed of
1628 C.R.M. Afonso et al. / Acta Biomaterialia 6 (2010) 1625–1629
 interplanar spacing strain mapping using geometrical phase analysis, which reveals the predominantly undistorted b matrix (strain around 0, between red and
Fig. 4. ð1 0 1Þb
green) and the shortened b0 phase domains (strain around 10%). (a) GPA ð1 0 1Þb interplanar spacing strain map combined with the original HR-TEM. (b) GPA ð1 0 1Þb
interplanar spacing strain map, revealing a large number of dislocations, the predominantly undistorted b matrix and b0 phase domains (dotted line highlighted region),
 interplanar spacing strain around 10%.
which present a ð1 0 1Þb
Table 1
Chemical compositions obtained by EDX nano-sized probe analysis (wt.%).
Region Ti Nb Ta Zr
b (dark) 51.3 ± 0.9 35.4 ± 0.8 5.8 ± 0.2 7.4 ± 0.5
b0 (bright) 50.4 ± 0.9 32.6 ± 0.8 7.1 ± 0.3 10.0 ± 0.5
GB ? a 53.8 ± 1.0 30.6 ± 0.7 6.5 ± 0.3 9.1 ± 0.5
Cu-rich platelets, coherent with the {1 0 0} Al matrix atomic plane
[9,10]. This phenomenon is well known in thin films and multilay-
ers, where incoherence between different layers can lead to con-
siderable softening [11–13]. Furthermore, martensite formed in
shape memory alloys does not result in an increase in the elastic
modulus because of the incoherency of the interface between the
martensite and a second phase [11]. Since the lattice parameters
of bcc (ab = 3.3 Å) and bct (ab0 = 3.0 and cb0 = 3.3 Å) phases are quite
similar in the alloy studied in this work, the interface between both
phases is likely to show a high degree of coherency, which results
in an elastic strain field along the structure and provokes an in-
crease in Vickers hardness.
Up to now there has been no microstructural evidence of phase
separation in stable b-Ti–Nb-based alloys containing high contents
of b-stabilizer elements such as Nb, Ta and Zr. Only supposition can
be found based on XRD and TEM analysis for other systems, how-
ever, the use of advanced microstructure characterization tech-
niques are fundamental to verify these suppositions. Li et al. [14]
and Moffat and Larbalestier [15] supposed that the separated
phase b0 (rich in b-stabilizer elements) acts as an a nucleus. Phase
separation in b-Ti alloys has been observed in other studies [11–
13], where the b0 phase was more easily detected, even by conven-
tional XRD analysis of Ti–Ta and Ti–Zr alloys with high contents of
b-stabilizer elements [12–16], whose bcc structures have a lattice
parameter comparable with b-Ti (aTi-b = 3.31, aNb = 3.30, aTa =
3.28, aZr = 3.57 Å). In Yang’s study of a 50Ti–30Zr–10Nb–10Ta
(at.%) alloy phase separation between b-Ti and b-Zr could be easily
detected by conventional XRD and, in the same way, there was a
hardness peak at around 400 °C for a sample isothermally aged
for 4 h [2], but no microstructural evidence of phase has been con-
firmed in the literature.

Long-term aging of Ti–35Nb–7Zr–5Ta alloy at 400 °C for 90 h
led to the formation of a metastable isothermal x phase, which
was easily detected by conventional XRD analysis [2]. Although
there was a considerable content of Zr (which retards x precipita-
tion) in the presently studied alloy it seems that the phase separa-
tion through spinodal decomposition might be a metastable state
that precedes the precipitation of a stable a phase. This fact was
confirmed by XRD analysis of long-term aged samples, where the
x phase peaks clearly appeared for heat treatments at 300 and
400 °C for 90 h [2]. It is well known in the literature that the x
phase acts as a nucleation site for the formation of a stable a phase
[17–19]. Another result that supports this is the hardness peak of a
sample aged at around 400 °C for 4 h, where the phase separation
occurred (a coherent interface was responsible for this hardening),
creating a strengthened crystalline lattice. Therefore, depending on
the aging conditions, the proposed precipitation sequence during
annealing would be: b ? b + b0 ? b + x ? b + a. If short aging
times are used, such as 4 h, adopted in this work, metastable x for-
mation might be occulted due to accelerated kinetics conditions,
resulting in b ! b þ b0 or b ! b þ a [19–21]. However,
ð400 C=4 hÞ ð600 C=4 hÞ
adopting a long-term aging time of 90 h [2] might result in slower
kinetics, suitable for precipitation of the x phase: b ! b þ x.
 ð400 C=90 hÞ
4. Conclusions
Under suitable aging conditions (400 °C for4 h) the Ti–35Nb–
7Zr–5Ta alloy underwent coherent spinodal decomposition of the
b phase into two solid solutions with different compositions and
elastic strains associated with domains of the Nb-rich b and Ta–
Zr-rich b0 phases. b is a body-centered cubic (bcc) phase
(ab = 3.3 Å) and b0 is a body-centered tetragonal (bct) phase with
a slight shortening in parameter ab0 = 3.0 Å compared with
cb0 = 3.3 Å. The small difference in (1 0 1) d spacings (db = 2.3 Å,
db0 = 2.1 Å) was confirmed by the splitting of spots on the Fourier
space map obtained from a HR-TEM image, evidence of the pres-
ence of a coherent interface between the separated phases (b and
b0 ), which explains the hardening of the alloy. Through GPA analy-
 C.R.M. Afonso et al. / Acta Biomaterialia 6 (2010) 1625–1629
sis, using an HR-TEM image, the localized strain map showed 5–
10 nm phase domains. Composition fluctuations of heavy ele-
ments, such as Ta and Zr, induce lattice parameter distortions
and, consequently, the splitting of peaks in the XRD pattern and
Fourier space map.
Acknowledgements
The authors would like to thank FAPESP, CAPES and CNPq for
financial support and LME/LNLS (Brazilian Synchrotron Light Labo-
ratory) for the HR-TEM analysis using a JEOL JEM 3010-URP micro-
scope and high resolution XRD experiments using synchrotron
radiation in the D10B-XPD (polycrystals X-ray diffraction)
beamline.
Appendix A. Figures with essential colour discrimination
Certain figures in this article, particularly Figure 4 is difficult to
interpret in black and white. The full colour images can be found in
the on-line version, at doi: 10.1016/j.actbio.2009.11.010).
References
[1] Tang X, Ahmed T, Rack HJ. Phase transformations in Ti–Nb–Ta and Ti–Nb–Ta–
Zr alloys. J Mater Sci 2000;35:1805–11.
[2] Ferrandini PL, Cardoso FF, Souza SA, Afonso CRM, Caram R. Aging response of
the Ti–35Nb–7Zr–5Ta and Ti–35Nb–7Ta alloys. J Alloys Compd
2007;433:207–10.
[3] Afonso CRM, Aleixo GT, Ramirez AJ, Caram R. Influence of cooling rate on
microstructure of Ti–Nb alloy for orthopedic implants. Mater Sci Eng C
2007;27:908–13.
[4] Yang G, Zhang T. Phase transformation and mechanical properties of the
Ti50Zr30Nb10Ta10 alloy with low modulus and biocompatible. J Alloys Compd
2005;392:291–4.
1629
[5] Ikeda M, Komatsu S-Y, Nakamura Y. Effects of Sn and Zr additions on phase
constitution and aging behavior of Ti-50mass%Ta alloys quenched from b
single phase region. Mater Trans 2004;45:1106–12.
[6] Hÿtch MJ, Snoeck E, Kilaas R. Quantitative measurement of displacement and
strain fields from HREM micrographs. Ultramicroscopy 1998;74:131–46.
[7] Stadelmann P. EMS – a software package for electron diffraction analysis and
HREM image simulation in materials science. Ultramicroscopy 1987;21:
131–45.
[8] Hÿtch MJ, Putaux J-M, Penisson J-M. Measurement of the displacement field of
dislocations to 0.03 angstrom by electron microscopy. Nature 2003;423:
270–3.
[9] Guinier A. Striated structure of age-hardened alloys. Nature 1945;155:569.
[10] Preston GD. Structure of age-hardened aluminium–copper alloys. Nature
1938;142:570.
[11] Söderberg H, Odén M, Molina-Aldareguia JM, Hultman L. Nanostructure
formation during deposition of TiN/SiNx nanomultilayer films by reactive
dual magnetron sputtering. J Appl Phys 2005;97:1143271.
[12] Van Steenberge N, Das J, Concustell A, Sort J, Suriñach S, Eckert J, Baró MD.
Influence of annealing on the microstructure and hardness of
Ti67.79Fe28.36Sn3.85 nanocomposite rods. Scripta Mater 2006;55:1087.
[13] Nakaniwa M, Sugimoto K, Kamei K, Nakamura Y, Sugimoto T. Effect of 2nd
phase on martensitic-transformation temperature in Cu–Al–Ni shape memory
alloys. J Jpn Inst Met 1993;57:135–57.
[14] Li S, Hao Y, Yang R, Cui Y, Niinomi M. Effect of Nb on microstructural
characteristics of Ti–Nb–Ta–Zr Alloy for biomedical applications. Mater Trans
2002;43:2964–9.
[15] Moffat DL, Larbalestier DC. The competition between martensite and omega in
quenched Ti–Nb alloys. Metall Trans A 1988;19:1677–86.
[16] Koul MK, Breedis F. Phase transformations in beta isomorphous titanium
alloys. Acta Metall 1970;18:579.
[17] Prima E, Vermaut P, Texier G, Ansel D, Gloriant T. Evidence of a nanophase
heterogeneous nucleation from x particles in a b-metastable Ti-based alloy by
high-resolution electron microscopy. Scripta Mater 2006;54:645–8.
[18] Narayanan GH, Archbold TF. Decomposition of metastable beta phase in all-
beta alloy Ti–13V–11Cr–3Al. Metall Trans 1970;1:2281.
[19] Zhou YL, Niinomi M, Akahori T. Effects of Ta content on Young’s modulus and
tensile properties of binary Ti–Ta alloys for biomedical applications. Mater Sci
Eng A 2004;371:283–90.
[20] Banerjee R, Nag S, Stechschulte J, Fraser HL. Strengthening mechanisms in Ti–
Nb–Zr–Ta and Ti–Mo–Zr–Fe orthopaedic alloys. Biomaterials 2004;25:3413–9.
[21] Banerjee R, Nag S, Fraser HL. A novel combination approach to the
development of beta titanium alloys orthopedic implants. Mater Sci Eng C
2005;25:282–9.